Reranker NVIDIA icon

Reranker NVIDIA

Use Nvidia Reranker to reorder documents after retrieval from a vector store by relevance to the given query.

Reranker NVIDIA

AI-generated

Summary

Rerank documents by relevance to a query using NVIDIA’s API. Connect this node as an AI Reranker; it returns up to Top N documents in ranked order, with each document’s metadata updated to include its relevance score.

Inputs

  • Model (required) — Model identifier used for reranking; defaults to nv-rerank-qa-mistral-4b:1.
  • Top N — Maximum number of ranked documents to return; defaults to 3.

Output shape

An AI Reranker connection that supplies a document compression function; given documents and a query, the function returns a list of ranked documents.

Sends the query text and each document’s page content to the configured NVIDIA endpoint. Results are ordered by the API ranking, limited to Top N, and each returned document receives metadata.relevanceScore from the API logit. Empty document lists return an empty list.

Links

Discussion